pydecimal does not properly handle positive keyword

[sshishov]

The same issue is happening for pydecimal as it was reported for pyfloat: #1928

I think that it should work more straightforward and more reliable, like: we should have a min and max values being updated and only at the end we generate the value.

MIN:

• min unset, positive unset, result = minimum calculated from places
• min set, positive unset, result = maximum of min and minimum calculated from places
• min unset, positive set to True, result = maximum between 0 and minimum calculated from places
• min unset, positive set to False, result = minimum calculated from places
• min set and positive set to True, result = maximum between 0 and min
• min set and positive set to False, result = maximum of min and minimum calculated from places

MAX:

• max unset, positive unset, result = maximum calculated from places
• max set, positive unset, result = minimum of max and maximum calculated from places
• max unset, positive set to True, result = maximum calculated from places
• max unset, positive set to False, result = minimum between 0 and maximum calculated from places
• max set and positive set to True, result = minimum of max and maximum calculated from places
• max set and positive set to False, result = minimum between 0 and max

Afterwards we should check min <= max and then return the random value from this range

I guess this approach is safe and reliable and will indicate about issues in usage when you try to build pydecimal from constrains which produce invalid range

[parsariyahi]

I was reading the source code and I saw that the same code in #1928 is used here

        if min_value is not None and min_value >= 0:
            sign = "+"
        elif max_value is not None and max_value <= 0:
            sign = "-"
        else:
            sign = "+" if positive else self.random_element(("+", "-"))

So, if we get to else statement the sign will choose randomly, and I think that's a bug.
But I was thinking about it and I didn't find any usage for a positive param in pydecimal fucntion.

So let's say we want negative numbers, we set max to 0 and the code should be functioning.
And other hand let's sat I don't want any negatives, so min to 0 and now I know I won't get any.
Or I don't care about the sign, and don't use any of min or max and the code should be random for +, -.